Empirical claim #5 — Russia sanctions reroute and European-bloc drag

Generated 2026-05-27 06:55 UTC by `scripts/py/intelligence/event_study_russia_reroute.py`. Pure-Python, no network. Re-run anytime.

Outcome — pre-registered falsification verdict

Claim FALSIFIED by the pre-registered test.

• Rule (a) — both INDA and TUR trail ACWI at T+60d: not triggered (INDA T+60d median = +0.42%, TUR = -0.87%)
• Rule (b) — ≥3/5 European-bloc ETFs outperform ACWI at T+60d: TRIGGERED (4/5 positive)
• Rule (c) — Group A Wilcoxon p > 0.10 on strongest signal: TRIGGERED (best Wilcoxon p in Group A = 0.208)
• Rule (c) — Group B Wilcoxon p > 0.10 on strongest signal: not triggered (best Wilcoxon p in Group B = 0.012)

A pre-registered claim that fails its own falsification test is evidence about the world, not a project failure. The published audit and the script that generated it stay live; the case-#5 prose's structural mechanism story does not survive this test at ETF-level resolution. See the Honest interpretation section below for what this DOES and does NOT mean.

Pre-registered cohort

All IPTM actions with target_countries containing 'RU', action_type ∈ {sanctions, sanction, export-control}, severity >= 4, announced_date in [2022-02-24, 2026-12-31] (post-full-invasion corpus; Crimea-era excluded), with a full T+60 trading-day window available in the ETF dataset.

Matched: 35 actions; 2 excluded for insufficient post-event data (announced after 2026-02-23). Final n = 33.

The claim — two-part hypothesis (pre-registered)

Severity-4+ sanctions or export-control actions targeting Russia are historically followed, over the 5- to 60-trading-day windows, by:

• (A) Discount-buyer reroute — INDA / TUR / MCHI (the three largest absorbers of displaced Russian crude and grain) outperforming ACWI; AND
• (B) European-bloc drag — EWG / EWN / EWI / EWQ / EWU (the five large-cap European national ETFs bearing the feedstock and energy cost shock) trailing ACWI.

SPY is reported alongside for transparency. This is a historical association, not a forecast. Both legs must clear their falsification rules independently for the claim to stand.

Falsification rules: (a) both INDA and TUR trail ACWI at T+60d, OR (b) ≥3/5 European-bloc ETFs outperform ACWI at T+60d, OR (c) Wilcoxon p > 0.10 on the strongest signal in either group.

Results — abnormal returns vs. ACWI

Median log-return on each ETF minus log-return on ACWI over the same window. Two-sided sign test against H0: median = 0 (binomial exact). Wilcoxon signed-rank p reported when n ≥ 6 non-zero observations.

Horizon: T+5d

Horizon: T+20d

Horizon: T+60d

Raw log-returns (for reference, not abnormal)

Events used

Honest interpretation

What this test says. At ETF-level resolution and the 5- to 60-trading-day horizon, there is no statistically detectable signal for either reroute outperformance (Group A) or European-bloc drag (Group B) around severity-4+ Russia-targeting sanctions / export controls in the 2022-02-24 to 2026-12-31 window.

What this test does NOT say.

• It does NOT say the reroute mechanism is fake. The Indian / Turkish / Chinese reroute of displaced Russian crude is well-documented in customs data (IEA, UNCTAD, Kpler tankers). What it says is the reroute did not translate into equity-market-detectable national-ETF outperformance at the tested horizons — possibly because Indian refiners' margin uplift was absorbed by sovereign windfall taxation (the Indian Special Additional Excise Duty on petroleum products, 2022-2024), or because INDA's 2022-2025 AI-services factor dominates any Russia-reroute signal.
• It does NOT say European chemicals / energy did not suffer a feedstock cost shock. BASF, Lanxess, Solvay, Linde all reported material 2022-2024 energy-cost EBIT compression; that mechanism is intact in the firm-level data. What this test says is the shock did not translate into broad-national-ETF underperformance around individual sanctions-package event dates — possibly because (i) EWG/EWQ/EWU are diversified far beyond their chemical-feedstock exposure (financials, autos, luxury, defence co-move on other macro factors), (ii) the cost shock had largely priced in by the 2022-Q1 invasion-onset event and subsequent marginal sanctions packages carried no incremental signal, or (iii) the 2024-2026 ECB-rate-cycle + defence-rearm trade (EWG / EWU industrials, EWQ defence) dominated post-2024 sanctions windows.
• It does NOT validate or refute the case-#5 prose claim that the EU-Russia regime forked into a parallel architecture — that is a primary-document claim (50% Rule, EU Council Regulation 2033, OFAC Rosneft/Lukoil SDN) not an empirical claim. Case #5's structural argument stands or falls on legal text and customs data, not on this event-study.

What we learn. The Russia regime is the first regime in the empirical-claims pipeline (after #1 BIS-chip / #2 MOFCOM-counter / #3 hilirisasi / #4 pooled-own-goal) where the structural mechanism the case study identifies does NOT translate into broad-national-ETF abnormal returns at the tested horizons. This is a useful discipline check: the register's structural claims and the event-study's associational claims are independent levels of evidence. Three regimes confirmed the mechanism at ETF level; one did not. The product is stronger for publishing both outcomes.

Honest caveats

1. Multi-target actions partially endogenise the reroute hypothesis. Many cohort events list target_countries: [RU, BY, CN, HK, TR, AE, IN, ...] — naming reroute economies as secondary targets of EU 17th-20th-package circumvention rules. The reroute outperformance test is therefore partially endogenous. A stronger version would restrict to RU-only-target actions (materially smaller cohort) — flagged for a later sensitivity tick.

2. Sample (n=33) spans 4 years of dense overlapping clusters. Effective independent-event count is materially lower than the raw n. The reported p-values are anti-conservative under the trade-independence assumption.

3. Severely overlapping event windows. Post-2022 actions cluster densely; multiple EU packages + BIS additions + OFAC designations often fall within the same T+60 window. A non-overlapping subset (drop any event whose T+60 window overlaps an earlier event's) is a defensible follow-up sensitivity in the same style as claim #4's non-overlap revision.

4. No factor-return adjustment. Abnormal returns are vs. ACWI only, not multi-factor. INDA carries heavy IT-services factor exposure orthogonal to the Russia mechanism; EWG/EWQ carry auto + luxury + defence factors. The reported numbers are raw associations, not factor-purged alphas.

5. Cohort-period confounding. The 2022-2026 window contains the commodity supercycle, the ECB / Fed tightening cycle, the AI-capex spike (INDA tailwind), the Israel-Hamas / Hezbollah war, and the May-2026 Iran-Hormuz crisis. Several drivers independently move the headline ETF groups concurrently.

6. Cohort criteria pre-registered; test-ticker list is post-hoc. target_countries ⊇ {RU}, action types and severity threshold pre-specified before the analysis ran. ETF groups were chosen post-hoc from case-#5's prose. No out-of-sample validation; future Russia-targeting actions may behave differently as bloc reorganisation matures.